Interpreting non-linguistic utterances by robots

Abstract

This paper presents a survey in which participants were asked to interpret non-linguistic utterances made by two different types of robot, one humanoid robot and one pet-like robot. The study set out to answer the question of whether the interpretation of emotions differed across types of robots, participant parameters and classes of utterance. We found that both male and female subjects were consistently more coherent in interpreting human over animal utterances, and animal over technological utterances. This held true with regard to the emotional and intentional interpretation, as well as the perception of appropriateness of a particular utterance with respect to a particular type of robot. We also found that males and females frequently differed significantly in their emotional and intentional interpretations of utterances. Finally, our results indicate that the morphology of a robot influences peoples judgment of what class of utterance is deemed appropriate for a particular type of robot.